Skip to main content
The Model Gateway at aisa.one/models is the live catalog of every model AIsa routes to — currently 59 across 9 providers. Filters let you slice the catalog by provider, capability, use case, and free-text search.

Open Model Gateway

Browse the live catalog with all filters available.

Compare models

Put two or three models side-by-side and share the result.

​
Filter types

Four filter sets combine (AND logic across sets, OR within a set) to narrow the result list.
1

Filter by Provider

Click any provider chip — Alibaba, Anthropic, ByteDance, DeepSeek, GLM, Google, MiniMax, Moonshot, or OpenAI. Each chip shows the count of models for that provider (e.g., Anthropic (15), OpenAI (12)).Toggle Multi-select (top-right of the provider row) to combine multiple providers.
2

Filter by Capability

Choose what the model must support: Text, Audio, Coding, Image, Video, Vision. Selecting a capability also reveals sub-capabilities below the main row — e.g., selecting Vision exposes Vision, Coding, Document, Spatial sub-filters you can stack for more specific matches.
3

Quick Presets

Curated one-click filters for common use cases:
  • Best for Coding — reasoning + coding models ranked for software tasks
  • Multimodal — text + vision + audio in one call
  • Deep Reasoning — models with strong chain-of-thought
  • Creative Writing — optimized for long-form generation
  • Translation — dedicated translation models
  • Image Generation — Seedream, Wan, Gemini image
  • Voice & Audio — speech-to-speech, audio understanding
  • Video — video generation and analysis
  • Budget-Friendly — lowest-cost tier across providers
4

Keyword Search

Free-text search at the top of the result list matches model names, providers, and endpoint paths. Examples: gpt-5, claude, anthropic, /v1/messages.

​
Saving a filter

Once you’ve applied a combination, click Save filter to the right of the active filter chips. Your saved filter is preserved as a URL so you can bookmark it or share it with a teammate. For example: aisa.one/models?tags=vision jumps straight to the 48 vision-capable models.

​
Result view

Each model card in the result grid shows:
FieldWhat it is
Provider logoAnthropic, OpenAI, Google, etc.
Model IDThe exact string to pass as model in API calls
EndpointWhich AIsa endpoint path serves this model
Input / OutputPrice per 1M tokens (or flat price for media models)
CapabilitiesText, Vision, Coding, etc. plus sub-capabilities
CompareClick to add this model to the comparison dock
Switch between grid and list view using the toggle to the right of the search bar. Sort results with the dropdown — options include Name (A–Z), Name (Z–A), Input Price (low → high), and Output Price (low → high).

​
Tips

The filter panel is stateful in the URL. Bookmark ?tags=coding,reasoning or ?preset=budget-friendly to jump back to a known slice.
To start over, click Clear next to any active filter chip or Clear all above the result grid.